obesity isn’t the problem

A few years ago, I was in the RWJ public health research seminar at Michigan, when I raised my hand and said something like this: “According to these tables, high BMI individuals are *not* at higher risk for mortality or morbidity. Maybe obesity isn’t the problem we think it is.” Silence ensued. And then, in a very British manner, the conversation just moved on, as if the little boy at the end of the table made a walrus face with the carrot sticks.

Turns out that I am not crazy. In Time magazine article, sociologist Abigail Saguy reviews recent research on weight and health and finds that BMI, and other measures of weight, are not well correlated, if at all, with health:

Yes, there are certain health risks associated with having an elevated BMI, such as type 2 diabetes and heart disease. More broadly, a higher BMI is associated with a greater risk of cardiometabolic abnormalities, as measured by blood pressure, triglycerides, cholesterol, glucose, insulin resistance and inflammation. Nonetheless, almost one quarter of “normal weight” people also have metabolic abnormalities, and more than half of “overweight” and almost one third of “obese” people have normal profiles, according to a 2008 study. That’s 16 million normal-weight Americans who have metabolic abnormalities and 20 million obese (or 56 million overweight and obese) Americans who have no such abnormalities.


If the AMA’s goal is to address the serious diseases of type 2 diabetes and heart disease, it would be more productive and accurate for the association to urge doctors to focus on cardiometabolic risk, recognizing that there are both metabolically healthy and metabolically unhealthy individuals in all categories of weight. Rather than promote weight loss per se, doctors should instead encourage their patients of all sizes to incorporate physical activity and a balanced diet into their lives.

In other words, doctors committed some serious selection bias errors when linking obesity with morbidity. Then, they simply ignored the evidence that was obvious from mortality tables.

Now, this doesn’t mean that one should ignore weight. Personally, I feel much better having lost some weight. I also engage in modest exercise. But the health policy point is that individual health simply isn’t correlated with weight. Instead, it’s about diet and cardiovascular health.

Adverts: From Black Power/Grad Skool Rulz

Written by fabiorojas

June 28, 2013 at 12:04 am

Posted in fabio, mere empirics

4 Responses

Subscribe to comments with RSS.

  1. Also, don’t athletes also often get counted as obese because of high (muscular) body weight? Did they parse these people out?



    June 28, 2013 at 1:02 am

  2. Over at crookedtimber there was a discussion awhile back about “The obesity paradox paradox” ( The author proposes that the criteria for “normal” and “obese” in terms of BMI was set too low and that if these criteria were raised then there would be stark rather than nearly equivalent difference of mortality rates for each group.



    June 28, 2013 at 1:18 am

  3. Fabio, go across campus and tell Victor Oladipo that his BMI shows him to be overweight.



    June 28, 2013 at 1:54 am

  4. I started a reply here, but it got so long I moved it over to scatterplot.



    June 29, 2013 at 2:17 pm

Comments are closed.

%d bloggers like this: